arXiv 2303.098062023-10-17cs.RO

DexRepNet：利用几何与空间手-物体表示学习灵巧机器人抓取网络

DexRepNet: Learning Dexterous Robotic Grasping Network with Geometric and Spatial Hand-Object Representations

College of Control Science and Engineering, Zhejiang University（浙江大学控制科学与工程学院）；Key Lab of CS&AUS of Zhejiang Province（浙江省计算机与应用系统安全重点实验室）；Department of Computer Science, National University of Singapore（新加坡国立大学计算机科学系）

Qingtao Liu, Yu Cui, Qi Ye, Zhengnan Sun, Haoming Li, Gaofeng Li, Lin Shao, Jiming Chen

备注：: accepted by IROS 2023

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

提出DexRep表示和DexRepNet方法，通过几何与空间手-物体交互特征提升灵巧抓取策略的泛化能力，在仿真和真实实验中取得高成功率。

摘要：

机器人灵巧抓取因多指机械手的高自由度（DoF）和复杂接触而成为一个具有挑战性的问题。现有的基于深度强化学习（DRL）的方法利用人类演示来降低因灵巧抓取的高维动作空间带来的样本复杂度。然而，对手-物体交互表示在高层次泛化方面的关注较少。本文提出了一种新颖的几何与空间手-物体交互表示，名为DexRep，用于在抓取过程中捕获动态物体形状特征以及手与物体之间的空间关系。DexRep由占用特征（Occupancy Feature）组成，用于通过移动手来感知范围内的粗略形状；表面特征（Surface Feature）用于变化的手-物体表面距离；局部几何特征（Local-Geo Feature）用于与潜在接触最相关的局部几何表面特征。基于这种新表示，我们提出了一种灵巧深度强化学习方法，以学习可泛化的抓取策略DexRepNet。实验结果表明，我们的方法在抓取成功率和收敛速度上均显著优于使用现有表示进行机器人抓取的基线方法。在已见物体上实现了93%的抓取成功率，在仿真和真实世界实验中，对未见类别的多种物体均实现了高于80%的抓取成功率。

英文摘要：

Robotic dexterous grasping is a challenging problem due to the high degree of freedom (DoF) and complex contacts of multi-fingered robotic hands. Existing deep reinforcement learning (DRL) based methods leverage human demonstrations to reduce sample complexity due to the high dimensional action space with dexterous grasping. However, less attention has been paid to hand-object interaction representations for high-level generalization. In this paper, we propose a novel geometric and spatial hand-object interaction representation, named DexRep, to capture dynamic object shape features and the spatial relations between hands and objects during grasping. DexRep comprises Occupancy Feature for rough shapes within sensing range by moving hands, Surface Feature for changing hand-object surface distances, and Local-Geo Feature for local geometric surface features most related to potential contacts. Based on the new representation, we propose a dexterous deep reinforcement learning method to learn a generalizable grasping policy DexRepNet. Experimental results show that our method outperforms baselines using existing representations for robotic grasping dramatically both in grasp success rate and convergence speed. It achieves a 93% grasping success rate on seen objects and higher than 80% grasping success rates on diverse objects of unseen categories in both simulation and real-world experiments.

机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。